Premature infants are a vulnerable population with multiple inter-related health problems that put them at risk for poor outcomes. Electronic health records capture large amounts of information that may help guide decisions, but existing alert and reminder-based clinical decision support (CDS) frameworks do not adequately apply multiple overlapping care guidelines to complex patient histories to produce coherent clinical recommendations. Our proposed study directly addresses the broad challenge area of 10 Information Technology for Processing Health Care Data for Research and the specific challenge topic 10-LM-102 Advanced Decision Support for Complex Clinical Decisions. This study will use a rules-based expert system embedded in an electronic health record (EHR) to extract, interpret, and present salient facts and recommendations related to the healthcare of premature infants. This proposal also meets the challenge of providing an immediate stimulus to the economy through the retention or creation of jobs and increased spending in the Delaware Valley region. The Children's Hospital of Philadelphia (CHOP) contributes substantially to the local economy. In 2008, CHOP's operations created and supported over 16,882 jobs in the region, and CHOP's total economic impact was over $2.01 billion. Moreover, through a combination of private donations, NIH funding, and allocations from its hospital operations, CHOP receives more total research support than any other children's hospital in the United States-$180 million in fiscal year 2007-2008. The current proposal will create or retain 5 jobs. The care of premature infants is a rapidly growing public health concern in the United States, with over 60,000 infants born in 2006 with a birth weight under 1500 grams1. With recent advances in neonatal care, more premature infants are surviving to discharge from the neonatal intensive care unit (NICU). Even though this high-risk group of patients is particularly sensitive to the care they receive after discharge, efforts to ensure high quality post-discharge care are haphazardly implemented. Electronic health records with embedded clinical decision support tools such as rules-based expert systems provide a natural opportunity to favorably affect the healthcare for these vulnerable infants. The complex decision-making for these infants must consider large numbers of variables that change over time. The matrices of rules required to cover all possible combinations of variables are huge. To date no such decision support tools have been implemented or evaluated to handle the complexity of decision-making required for the healthcare of premature infants. This project will use a prospective cluster randomized design to evaluate the success of a health information technology intervention to improve care quality for low birth weight (LBW) and very low birth weight (VLBW) premature infants from the time of intensive care nursery discharge through 24 months of age. The intervention will consist of (1) a real-time data mining tool functioning in the EHR that will organize large amounts of health information visually in a succinct time-line format;(2) a rules-based expert system that will forecast a schedule for upcoming preventive health assessments and interventions appropriate for LBW and VLBW infants, and (3) a suite of problem focused expert systems to guide decision-making related to the most prevalent co-morbidities and complications experienced by this vulnerable population. Evaluation will focus on three areas: (1) usability of the automated data-mining tool and expert system pre-implementation;(2) change in clinician knowledge of standard care process for LBW and VLBW infants pre- and post-intervention;and (3) change in care process outcomes in the intervention practices compared to the control practices. The proposed work will improve the capacity for delivering clinical decision support (CDS) in complex clinical domains. Rigorous evaluation of both the design phase and intervention phase will support publications regarding best practices for future CDS development. The approach of a standards-based web-service model with a modest custom programming "footprint" in the electronic health record (EHR) will facilitate the distribution of this and future CDS interventions to other healthcare organizations and EHR vendors. The proposed work will improve the capacity for delivering clinical decision support (CDS) in complex clinical domains. Rigorous evaluation of both the design phase and intervention phase will support publications regarding best practices for future CDS development. The approach of a standards-based web-service model with a modest custom programming "footprint" in the electronic health record (EHR) will facilitate the distribution of this and future CDS interventions to other healthcare organizations and EHR vendors.